Emerging patterns of CO2 : O2 dynamics in rivers and their link to ecosystem carbon processing

Aquatic metabolism is reflected in the dynamics of dissolved oxygen (O<inf>2</inf>) and carbon dioxide (CO<inf>2</inf>) concentrations. Thus, paired measurements of CO<inf>2</inf> and O<inf>2</inf> concentrations can capture the metabolic characteristics of an ecosystem, with promising results in lakes. Yet, for rivers, hydrological, chemical, and biological processes all influence CO<inf>2</inf> and O<inf>2</inf> concentrations, complicating how paired measurements can be used to infer ecosystem processes. Here we combine a data synthesis with a simple mechanistic model of river metabolism, gas exchange, groundwater inputs and carbonate equilibrium to assess how each imprints upon CO<inf>2</inf> : O<inf>2</inf> patterns. Among the physicochemical processes considered, groundwater inputs substantially influenced CO<inf>2</inf> : O<inf>2</inf> relationships. Regardless, analysis of paired CO<inf>2</inf> : O<inf>2</inf> data resolved predictable differences in ecosystem function across rivers with variable productivity and disturbance, as well as along the river continuum. Results indicate that paired CO<inf>2</inf> : O<inf>2</inf> data can aid in assessments of river metabolism, provided that we account for the dynamic physical environment.

We are thankful to the StreamPULSE Network, with funding provided by the National Science Foundation Macrosystems program (Grant EF-1442439), for openly providing the data supporting this study. We thank Sam Blackburn and Emily Stanley for the collection of data in Brewery and Black Earth creek, William McDowell for the collection of data in Saddleback mountain stream, and the Krycklan Catchment Study for data in Krycklan C6. We thank Rafael Marcé and Edward Stets for discussions on carbonate equilibrium, Núria Catalan for early comments on this article, and Dominic Vachon for the discussions that sparked this work. Finally, we appreciate the comments and criticisms from two anonymous reviewers who improved this work. Gerard Rocher-Ros was supported by a grant from the Swedish research Council (#2021-06667) and the European Research Council (ARIMETH-101161308). Susana Bernal and Xavier Peñarroya work was supported by the projects EVASIONA (PID2021-122817-NB-100) and RIPAMED (CNS2023-144737) funded by MICIU/AEI/10.13039/501100011033, FEDER, UE and Next GenerationEU/PRTR. Anna Lupon work was supported by the projects INHOT (CNS2022-135690) and FLUPRINT (EUR2023-143456), funded by MICIU/AEI/10.13039/501100011033 and NextGenerationEU/PRTR. Carolina Jativa was funded by the FPU21/03523 from MICIU/AEI/10.13039/501100011033 and FSE+.

Peer reviewed